Fast H-ARQ acknowledgement generation method using a stopping rule for turbo decoding
AssigneeInterDigital Technology Corporation
InventorsSung-Hyuk Shin, Nader Bolourchi
Life of this patent
Estimating the term…
Abstract
A stopping rule for Turbo decoding that is applied for both good and bad code blocks is disclosed. If the iteration either converges or diverges, decoding is terminated. In an alternative embodiment, the result of the stopping rule testing may be used for H-ARQ acknowledgement generation: if the iteration converges, an ACK is generated and if the iteration diverges, a NACK is generated. Optionally, the maximum number of decoding iterations may be dynamically selected based on MCS levels.
Claims36
A method employed in conjunction with turbo decoding for reducing a number of iterations, comprising:
- a) initializing an iteration count upon receipt of a turbo code block;
- b) increasing the iteration count responsive to performance of step (a);
- c) performing an iteration;
- d) increasing the iteration count responsive to performance of step (c);
- e) determining if a latest iteration converges;
- f) generating an acknowledgement responsive to step (e);
- and
- g) terminating performance of further iterations.
A method employed in conjunction with turbo decoding for reducing a number of iterations, comprising:
- a) initializing an iteration count upon receipt of a turbo code block;
- b) increasing the iteration count responsive to performance of step (a);
- c) performing an iteration;
- d) increasing the iteration count responsive to performance of step (c);
- e) determining if a latest iteration diverges;
- f) generating an acknowledgement responsive to step (e);
- and
- g) terminating performance of further iterations.
A method employed in conjunction with turbo decoding iterations, comprising:
- a) initializing an iteration count upon receipt of a turbo code block;
- b) increasing the iteration count responsive to performance of step (a);
- c) performing an iteration;
- d) increasing the iteration count responsive to performance of step (c);
- e) determining if a latest iteration converges;
- f) determining if the latest iteration diverges in the event that there was not convergence at step (e);
- g) generating an acknowledgement responsive to step (f);
- and
- h) terminating performance of further iterations.
A method employed in conjunction with turbo decoding for reducing a number of iterations, comprising:
- a) generating log-likelihood ratios (LLRs) based on a systematic bit sequence and a parity bit 1 sequence of inputted coded data and a priori data;
- b) permuting the LLRs;
- c) generating extrinsic information based on the permuted LLRs and a parity 2 sequence of the coded input data;
- d) de-interleaving the extrinsic information;
- and
- e) providing the de-interleaved extrinsic information as a priori data employed at step (h).
Apparatus employed in turbo decoding for reducing a number of decoding iterations, comprising:
- means for initializing an iteration count upon receipt of a turbo code block;
- means for increasing the iteration count responsive to said initializing means;
- means for performing an iteration;
- means for increasing the iteration count responsive to said performing means;
- means for determining if a latest iteration converges;
- means for generating an acknowledgement responsive to said determining means;
- and
- means for terminating performance of further iterations.
Apparatus employed in turbo decoding for reducing a number of decoding iterations, comprising:
- means for initializing an iteration count upon receipt of a turbo code block;
- means for increasing the iteration count responsive to said initializing means;
- means for performing an iteration;
- means for increasing the iteration count responsive to said performing means;
- means for determining if a latest iteration diverges;
- means for generating an acknowledgement responsive said means for generating an acknowledgement;
- and
- means for terminating performance of further iterations.
Apparatus for performing turbo decoding iterations, comprising:
- means for initializing an iteration count upon receipt of a turbo code block;
- means for increasing the iteration count responsive to performance of said initializing means;
- means for performing a decoding iteration;
- means for increasing the iteration count responsive to said performing means;
- means for determining if a latest iteration converges;
- means for determining if the latest iteration diverges in the event that convergence was not detected by said convergence determining means.
- means for generating an acknowledgement responsive to said divergence determining means;
- and
- means for terminating performance of further iterations responsive to said acknowledge generating means.
Apparatus for turbo decoding which reduces a number of decoding iterations, comprising:
- means for generating log-likelihood ratios (LLRs) based on a systematic bit sequence and a parity bit 1 sequence of inputted coded data and a priori data;
- means for permuting the LLRs;
- means for generating extrinsic information based on the permuted LLRs and a parity 2 sequence of the coded input data;
- means for de-interleaving the extrinsic information;
- and
- means for providing the de-interleaved extrinsic information as a priori data to said means for generating LLRs.
Dependencies are inferred from the claim text.